Expires: January 24, 2025
Anja Roden, M.D.
Professor of Medicine and Laboratory Medicine and Pathology
Director, Immunostains Laboratory
Division of Anatomic Pathology
Mayo Clinic, Rochester, Minnesota
Good afternoon. Today, I will give you an update on the 2021 World Health Organization (WHO) classification and ancillary testing in pleural mesothelioma.
I'm Anja Roden. I'm one of the thoracic pathologists at Mayo Clinic in Rochester, and I'm also the medical director of the Immunostains Lab.
I have nothing to disclose.
We will discuss the changes to the 2021 WHO classification. We will focus on mesothelioma in situ and diffuse mesothelioma, and then we'll talk about ancillary testing in mesothelioma.
As you know, pleural mesothelioma is a rare disease. However, if we look at the data from 2020, there were over 30,000 new cases of pleural mesothelioma worldwide, and almost as many patients died from the disease, so it is a very serious, deadly disease.
Looking at the age standardized rate of mesothelioma in the world, you can appreciate that most of the cases occur in North America, some of the European countries, South Africa, and Australia.
Mesothelioma has been associated — or the overall survival of mesothelioma has been associated — with the histologic subtype, with epithelioid mesotheliomas having the best overall survival, while sarcomatoid mesotheliomas have the worst survival. Survival also depends on stage, male gender, lymph node dissection, nuclear grading in the epithelioid subtype, and treatment without chemotherapy.
In the 2021 WHO classification, the word “malignant” has been removed for diffuse and localized mesotheliomas, as they all are malignant. And to distinguish them from the well-differentiated papillary mesotheliomas, these tumors now are coined well-differentiated papillary mesothelial tumors. We do focus more on architectural patterns, cytological and stromal features of mesotheliomas, and we are also now recommended to grade the epithelioid mesotheliomas. Mesothelioma in situ has been introduced to the 2021 WHO classification.
Let’s start with mesothelioma in situ.
This is thought to be a preinvasive lesion, which is represented by recurrent, unexplained pleural effusion. By definition, there should be no mass lesion on imaging studies or by thoracoscopy, and it is categorized by a single surface layer of neoplastic mesothelial cells, and I will show you an example shortly. These mesothelial cells should show either lack of expression of BAP1, and/or mTAP, or show homozygous deletion of CDKN2A. These lesions are not diagnosed on small biopsies, cytology, or effusions, and we really need multiple biopsies from different areas to make that diagnosis.
Here's one of such an example. You can appreciate the single layer of pretty bland appearing epithelioid mesothelial cells. However, they do show loss of BAP1 expression. In these loss of expression markers, we have to make sure that we do have a positive internal control, which is in this case, the BAP1 expression of benign endothelial cells. If we do not see this internal control, the stain might have been a random staining failure and cannot be interpreted.
Diffuse mesotheliomas usually represent with unilateral plural effusion. There can be invasion of contiguous structures, such as the pericardium or transdiaphragmatic, for instance, into the liver or the spleen. Lymphangitic involvement is more commonly seen with the epithelioid subtype and mediastinal lymph node involvement can occur, and later on in the disease, other lymph nodes may also be involved. Metastases — although not really common — they can also occur specifically to the lung, bone, liver, and brain.
Here's an example of a diffused mesothelioma, and we recognize the mesothelioma along the pleural surface, but also involving the fissures of the lung. On low power, here's the lung, and you can appreciate the mesothelioma spreading along the mesothelial surface or the pleural surface.
This is a biopsy of a left supraclavicular lymph node of a 59-year-old female who presented with pleural effusion and mediastinal lymphadenopathy. You can see that the lymph node is almost completely taken over by these nests and sheets of epithelioid cells. On high power, these cells are pretty bland. They have round nuclei with prominent nucleoli. However, they do stain with markers of mesothelial differentiation, such as CK5 and WT-1, and they were negative for carcinoma markers such as PAX8, TTF1, and pCEA. With that, this was diagnosed as metastatic mesothelioma epithelioid subtype.
The 2021 WHO classification of diffuse mesothelioma kept the subtypes of epithelioid, biphasic, or sarcomatoid morphology. And now, we have to pay more attention to the architectural patterns and the cytologic and stromal features, because some of those have prognostic relevance and pretty much all of them are important to avoid important differential diagnostic pitfalls. In resection specimens, we are supposed to stage the mesothelioma according to the TNM, and right now we stage according to the 8TH AJCC/UICC staging system.
In epithelioid mesotheliomas, the patterns include tubulopapillary, trabecular, micropapillary, solid, and adenomatoid patterns. Cytological features include clear cell, rhabdoid, deciduoid, small tumor cells, signet ring, pleomorphic, and lymphohistiocytoid features. Pleomorphic and lymphohistiocytoid features can also occur in sarcomatoid mesotheliomas, and in these cases, the distinction between epithelioid and sarcomatoid mesothelioma should be really based on the remaining tumor cell morphology. Uncommonly — but it can occur — we may see myxoid stroma in epithelioid mesotheliomas. All these features should be reported in biopsies and resection specimens of diffuse mesothelioma, and in resection specimens, ideally, we should report the percentages of these features.
Here are some examples of the architectural patterns. On the left side, you can see the tubulo-papillary pattern that is characterized by these papillae lined by the neoplastic cells, and they do have fibrovascular cores. On the right side is a trabecular pattern that is comprised of these single filing or dual filings of tumor cells.
Here's a micropapillary pattern, which is comprised of papillae which lack the fibrovascular cores. This is a solid pattern, and on the right side is an adenomatoid pattern characterized by these atypical gland-like structures that are lined by neoplastic mesothelial cells.
This picture on the right shows clear cell cytological features, and certainly these cells can come into the differential diagnosis of, for instance, metastatic renal cell carcinoma, clear cell type, or even some clear cell features in squamous cell carcinomas. In the middle, these are rhabdoid features, which are usually larger cells that have ample cytoplasm and eccentric nuclei with prominent nucleoli. The deciduoid cell features really remind us of the decidua of the placenta and is characterized by ample cytoplasm. The nuclei are round to oval, with inconspicuous or more conspicuous nucleoli and open chromatin.
In diffuse mesotheliomas, the epithelioid subtype should be graded, and that should occur on biopsies as well as resection specimens. The grading starts with scoring of the nuclear atypia, and if the nuclear atypia is only mild, that equals a score 1; moderate atypia equals a score 2; and severe atypia is a score 3. We also score mitotic activity and we count them per two square millimeters. A score 1 equals mitoses number one, or no mitoses. If you have two to four mitoses, this is a score 2. And five and higher mitotic counts are score 3. Now, we build a sum of the score of the nuclear atypia and a score for the mitotic count, and if the sum is a two or three, then this is a nuclear grade I. If the sum is a four or five, this is a nuclear grade II. And if the sum is a six, then this is a nuclear grade III.
We also have to look for necrosis and then we form the overall tumor grade. If the nuclear grade is I, or the nuclear grade is II without necrosis, then this equals a low overall tumor grade. If we have a nuclear grade II with necrosis or a nuclear grade III, this equals an overall high tumor grade.
Here are some examples. We really see very mild nuclear atypia. The nuclei are almost monotonous. They do not show prominent nucleoli and we do have ample cytoplasm. So this is a nuclear atypia score 1. There were no mitoses so there's a score 1 for mitotic activity, and therefore the sum score is two, which equals a nuclear grade I. There was also no necrosis and therefore the overall tumor grade is low.
This is in contrast to that tumor, which shows tumor nuclei that have prominent nucleoli, some of the tumor cells have multiple nuclei and many of the tumor cells have not much of cytoplasm.
There's also some atypical mitotic activity here. So, overall the nuclear atypia is a score 3. There were five mitoses per two square millimeters, which also equals a score 3. So the sum score is six, which equals a nuclear grade III. There was no necrosis, but because of the nuclear grade III, the overall tumor grade is high.
The grading should occur in areas of highest grade. Again, this should be reported in biopsies and resection specimens of diffuse mesotheliomas of epithelioid subtype. And this is because the grading is strongly predictive of overall survival in multivariate analysis.
Sarcomatoid mesotheliomas are categorized by these elongated spindle cells and they can be quite bland or very atypical. New to the WHO 2021 is the transitional cytologic feature and we will talk about that shortly. Other cytologic features include pleomorphic and lymphohistiocytoid features, and stromal features include desmoplastic features and heterologous elements.
Transitional cytologic features are basically features that are in between epithelioid and sarcomatoid mesothelioma cytologic features. So the transitional cytologic features are characterized by elongated and plump cells. They have usually a moderate amount of cytoplasm and prominent nucleoli as you can appreciate here. They do usually grow in sheet-like areas and, if in doubt, a reticulin stain might help.
How does reticulin stain help? Here on the left side, we have the transitional features and in those cases the reticulin stain will highlight single tumor cells. And that is in contrast to epithelioid subtypes, which we have here on the right side and in those cases, the reticulin stain builds around nests of tumor cells.
There are multiple manuscripts and studies on transitional cytologic features. One of the largest was from Galateau-Salle and colleagues, which looked at almost 50 cases of mesothelioma that had at least a transitional component and compared those to sarcomatoid and epithelioid mesotheliomas. The cases with a transitional component had a median survival of 6.7 months, which is similar to the sarcomatoid mesotheliomas. Also the rates of loss of BAP1 and homozygous CDKNA2 deletion in these cases was very similar to sarcomatoid mesotheliomas. They performed transcriptomic analysis and showed that these cases with transitional component clustered with sarcomatoid mesotheliomas and away from the epithelioid mesotheliomas. And because of all this data, the WHO felt that these lesions should be really subtyped together with the sarcomatoid subtype and not the epithelioid subtype.
So what are the emerging biomarkers?
Well, there's always difficulties with the distinction between malignant and reactive mesothelial proliferations. If we see invasion of the mesothelial cells in adipose tissue, lung and skeletal muscle, and the keratin stain might even help with that, then a malignancy is favored. Tumefactive growth pattern and haphazard growth pattern of neoplastic cells is also consistent with malignant proliferation in contrast to reactive. However, on small biopsies, we often do not have adipose tissue, lung, or skeletal muscle to even look for invasion, and sometimes it's also difficult to identify tumefactive growth pattern or have haphazard growth pattern on small biopsies. And in these conditions, loss of BAP1 and/or mTAP expression really comes in handy and is consistent with malignant proliferation. Homozygous deletion of CDKNA2 and or NF2 is also consistent with malignant proliferation and has not been described in reactive mesothelial proliferations. The expression of GLUT-1, p53, desmin, EMA, IMP3, and EZH2, although they have been proposed to be helpful in that distinction, they have insufficient specificity and should not be used for the distinction between malignant and reactive mesothelial proliferations.
Here's an example of a 77-year-old gentleman who presented with recurrent right plural effusion and underwent excision of the right parietal pleura. We do appreciate a thickened pleura and maybe there is some invasion into adipose tissue here. However, most of the biopsy appears to be hypocellular.
On high power, there are these atypical spindle cells. However, these spindle cells are pretty bland. It does appear to be invading into the adipose tissue, and as I said, sometimes a keratin stain can be really helpful because that really shows the invasion of these atypical cells into adipose tissue and that is really consistent with a malignant process.
Additional immunostains showed that these spindle cells are positive for calretinin. Some are at least positive for WT-1, and they are diffusely positive for GATA3 while being negative for markers of carcinoma including MOC31 and BerEP4. In addition, these spindle cells are negative for BAP1, so they lost expression of BAP1 with a nice internal control of BAP1 staining here in the endothelial cells. So this is mesothelioma sarcomatoid type.
This is a 53-year-old woman who presented with abdominal distension, and CT imaging studies revealed ascites and omental nodularity and the patient underwent paracentesis. The paracentesis was focally cellular, and within these areas we had clusters of atypical epithelioid cells with prominent nucleoli and these clusters stained positive for WT-1 and calretinin and CK5/6 and were negative for markers of carcinoma, including CD15, ER, and BerEP4. Now these cases used to be called atypical mesothelial proliferation or suspicious. However, now with BAP showing that these cells actually showed loss of BAP1 expression with a nice internal control, we were able to call this mesothelioma epithelioid type. Otherwise, the diagnosis of mesothelioma could only have been done 11 months later when the patient presented with a lesion in the inguinal region.
This is just an example, how a FISH for CDKN2A would look like if we have homozygous deletion of CDKN2A. We do appreciate two green dots; however, the two red dots are missing.
Now while the specificity for malignancy of loss of nuclear BAP1 expression and loss of cytoplasmic mTAP expression or homozygous CDKN2A deletion is 100%, the sensitivity of these markers really depends on the morphology. For instance, loss of nuclear BAP1 expression is more commonly seen in epithelioid mesotheliomas, while the loss of cytoplasmic mTAP expression and the homozygous CDKN2A deletion are more commonly seen in sarcomatoid mesotheliomas. So in practice, we usually use BAP1 and mTAP together as IHC markers, because it also has a faster turnaround time than the FISH studies for CDKN2A, and with using BAP1 and mTAP, we have a higher sensitivity than using each marker alone.11-16
There are some pitfalls to these markers because these markers are not useful in the distinction between mesothelioma and carcinoma as loss of BAP1 expression and mTAP expression and homozygous deletion of CDKN2A can be seen in various carcinomas. So these markers are only useful for the distinction between the reactive mesothelial proliferation and malignant mesothelial proliferation, but are not useful in distinction between mesothelioma and carcinoma.
Now for IHC markers to distinguish between mesothelioma and carcinoma, none of these markers are perfect, and therefore, we have to always use a panel of at least two markers for mesothelial differentiation and two carcinoma markers. Which markers we choose usually depends a bit on the differential diagnosis and the origin of the potential carcinoma. Good mesothelioma markers include WT-1, calretinin, D2-40, CK5, and recently GATA3 if diffusely expressed specifically in sarcomatoid mesotheliomas has been also found to be very useful after we exclude the possibility of urothelial carcinoma and breast carcinoma. In regards to carcinoma markers, pCEA, TTF-1, Napsin A, P40, PAX8 are useful, although in PAX8, we have to be cautious as some of the peritoneal mesotheliomas can express PAX8. I also try to stay away from MOC31 and BerEP4 because although they are carcinoma markers, they often stain at least focally positive in mesotheliomas. So if we can, we try to stay away from them. And keratin, such as AE1/AE3, CAM5.2, and CK7 are usually positive in both mesothelioma and carcinoma.
So in summary, we do need to distinguish diffuse mesothelioma from localized mesothelioma, mesothelioma in situ, well differentiated papillary mesothelial tumor, and reactive mesothelial proliferations, as there are significant differences in prognosis and management of these patients. We do need to grade diffuse epithelioid mesotheliomas in biopsies and resection specimens. And for the distinction of mesothelioma and carcinoma and of course, other malignancies such as mesotheliomas and sarcomas, we do need to use a panel of immunostains as none of the immunostains is specific for mesothelioma. Also, the distinction between reactive and malignant mesothelial proliferation is down to morphology, but immunostains and FISH may be helpful in that distinction.
And with that, I would like to thank you for your attention. Please email me or call me if you have any questions. And if you are interested in any references, here are the references that I commented on today.
Contact us: firstname.lastname@example.org